lable at ScienceDirect

Quaternary Science Reviews 29 (2010) 3875e3886

Contents lists avai

Quaternary Science Reviews

journal homepage: www.elsevier .com/locate/quascirev

Phase relationships of North Atlantic ice-rafted debris and surface-deep climateproxies during the last glacial period

David A. Hodell a,*, Helen F. Evans b, James E.T. Channell c, Jason H. Curtis c

aGodwin Laboratory for Paleoclimate Research, Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UKbBorehole Research Group, Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, NY 10964, USAcDepartment of Geological Sciences, University of Florida, P.O. Box 112120, Gainesville, FL 32611, USA

a r t i c l e i n f o

Article history:Received 15 January 2010Received in revised form28 August 2010Accepted 8 September 2010

* Corresponding author. Tel.: þ44 0 1223 330270;E-mail address: dah73@cam.ac.uk (D.A. Hodell).

0277-3791/$ e see front matter Crown Copyright � 2doi:10.1016/j.quascirev.2010.09.006

a b s t r a c t

We report stable isotope, core scanning XRF, and ice-rafted detritus (IRD) data in glacial-aged sedimentsfrom piston Core KN166-14-JPC-13 (hereafter referred to as JPC-13) retrieved at 53.1�N 33.5�W on thesouthernGardarDrift, NorthAtlantic. A chronologywas establishedbycorrelatingmillennial-scale featuresin the benthic d18O record to Portuguese Margin Core MD95-2042. Once the alignment of benthic d18O ofJPC-13 to MD95-2042 is fixed, the relative timing of proxy variables is used to determine the phasing ofchanges in IRD and surface-deep hydrography. Each peak inNorthAtlantic IRD coincidedwith a decrease inbenthic d18O that, in turn, has been linked to warming in Antarctica (Events A1eA7). The IRD pulses arefollowed shortly thereafter by abrupt decreases in planktonic d18O, indicating warming and increased heattransport to the subpolar North Atlantic associated with Greenland interstadials (GIS) 8, 12,14, and 16e17.Grain-size and elemental (K/Ti) variations indicate that IcelandeScotland Overflow Water (ISOW) wasstronger during these long, warm interstadial periods of Marine Isotope Stage (MIS) 3. The results areconsistent with interhemispheric phase relationships inferred from Iberian Margin sediment cores andwith methane synchronization of Greenland and Antarctic ice cores.

Crown Copyright � 2010 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Millennial-scale climate variability during the last glacial periodhas a somewhat different character between Greenland andAntarctica. Greenland ice cores display abrupt stadialeinterstadialoscillations resembling a ‘square-wave’ form (DansgaardeOeschger(DeO) events), whereas Antarctic ice cores show a ‘triangular’ formconsisting of Antarctic Isotope Maxima [AIM] that are accompaniedby variations in atmospheric CO2 (Indermühle et al., 2000; Ahn andBrook, 2008). The Greenland DeO events are bundled into longer-term packages, with the beginning of the cycle marked by warmingand the end of the cyclemarked byHeinrich events (Heinrich,1988;Bond et al., 1993). This pattern, sometimes referred to as a Bondcycle, is similar in time scale to the major features of Antarcticclimate variability albeit phase shifted.

Studying the phase relationships of millennial-scale variabilityis important for determining the relative timing among variousprocesses in the climate system. The interhemispheric phasing forthe last 90 kyrs has been determined by synchronizing methane

fax: þ44 0 1223 333450.

010 Published by Elsevier Ltd. All

records between Greenland and Antarctic ice cores (Blunier andBrook, 2001; EPICA Community Members, 2006). Results indicatethat the warm events in Antarctica coincided with the longest,coldest stadials in Greenland associated with Heinrich events. TheHeinrich stadials were immediately followed by abrupt warming inGreenland resulting in a long interstadial period at the same time asAntarctica began to cool. The phasing between Greenland andAntarctica has been described as either anti-phase (“seesaw”) orphase shifted (“Southern lead”), although neither descriptionadequately captures the complexity of the phase relationship (Steigand Alley, 2002). The out-of-phase behavior has been attributed tothe “bipolar seesaw” whereby changes in heat transport related toAtlantic Meridional Overturning Circulation (AMOC) results ina warming (cooling) in the north and cooling (warming) in thesouth when AMOC is strong (weak) (Mix et al., 1986; Crowley,1992;Broecker, 1998; Stocker, 1998; Stocker and Johnsen, 2003; Knuttiet al., 2004). The amplitude of Antarctic warm events showsa linear relationship with the duration of the accompanying stadialperiods in Greenland during MIS3 (EPICA Community Members,2006).

Determining phase relationships of millennial-scale variabilityin glacial periods beyond the last climate cycle is challengingbecause Greenland ice cores are presently limited to the last

rights reserved.

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e38863876

124 kyr. Absolute dating of marine cores is too imprecise tocorrelate and resolve small differences in the timing of paleo-climate signals (Andrews et al., 1999; Wunsch, 2006). An alterna-tive approach is to determine the relative phasing of changes inproxy variables in the same core thatmonitor different componentsof the ocean-climate system. Despite the merits of inferring phaserelationships in a single core, caution is needed because bio-turbation combined with changes in species abundances andsedimentation rates can affect lead/lag relationships among proxyvariables in the same core (Löwemark and Grootes, 2004;Löwemark et al., 2008).

One of the first studies to apply this technique was Charles et al.(1996) who compared planktonic d18O and benthic d13C in SouthAtlantic Core RC11-83 and concluded that northern hemisphereclimate change lagged those of the southern hemisphere by 1500years. This result was later confirmed by methane synchronizationof polar ice cores (Blunier et al., 1998, Blunier and Brook, 2001;EPICA Community Members, 2006).

Shackleton et al. (2000, 2004) produced a detailed planktonicd18O record of millennial-scale variability in Portuguese MarginCore MD95-2042 that can be confidently correlated to Greenlandice cores. The benthic d18O signal in the same core resemblestemperature records from Antarctic ice cores, although the cause ofthis correlation is not well understood. Changes in benthic d18Omay in large part reflect local deep-water d18Odw, related tochanges in deep-water sourcing and/or source signature (Skinneret al., 2007). Shackleton et al. (2000, 2004) correlated theplanktic d18O of MD95-2042 to Greenland and used methanesynchronization of polar ice cores to evaluate the phase relation-ship with Antarctica (Blunier et al., 1998; Blunier and Brook, 2001;EPICA Community Members, 2006). They found the relative timingof the planktonic and benthic d18O signals has the same inter-hemispheric phasing as that deduced from other studies (Blunieret al., 1998, Blunier and Brook, 2001); that is, millennial-scale

Fig. 1. Site location of Cores KN1666-14-JPC-13 (IODP Site U1304), MD95-2042, and IODP Sitrepresent a schematic illustration of the major components of the Atlantic Meridional Overtof the IRD belt during the last glacial period (Ruddiman, 1977). Red lines representISOW ¼ IcelandeScotland OverflowWater; DSOW ¼ Denmark Strait OverflowWater; WBUCAtlantic Current. (For interpretation of the references to colour in this figure legend, the re

warmings in Antarctica preceded the onset of Greenland warmingsand the onset of rapid warming in Greenland coincided withcooling in Antarctica (Charles et al., 1996; Blunier et al., 1998,Blunier and Brook, 2001; EPICA Community Members, 2006).

Following the lead of previous investigators (Charles et al., 1996;Shackleton et al., 2000), we applied a similar strategy to studyphase relationships of multi-proxy data from piston core KN166-14-JPC-13 (hereafter referred to as JPC-13) located on the southernGardar Drift in the subpolar North Atlantic (Fig. 1). A precisechronology (SFCP04) was established by correlating millennial-scale features in the benthic d18O record of JPC-13 to PortugueseMargin Core MD95-2042 (Fig. 2) that, in turn, has been synchro-nized with Greenland and Antarctic ice cores (Shackleton et al.,2004). By comparing the relative timing of proxy variables inJPC-13 relative to the benthic d18O signal, we infer the phasing ofchanges in surface and bottom water hydrography and their rela-tion to delivery of ice-rafted debris (IRD) to the subpolar NorthAtlantic. The focus of the paper is on the last 80 ka because this isthe period for which the common SFCP04 time scale is available forIberian Margin Core MD95-2042 and ice cores from Greenland(GRIP) and Antarctica (Vostok) (Shackleton et al., 2004).

2. Site location and hydrography

A 23.6-m piston core (KN166-14-13JPC) was recovered in 2002from the southernmost Gardar Drift near the Charlie Gibbs FractureZone (53� 3.410, 33� 31.780, 3082 m). The Gardar Drift is an elon-gated contourite deposited along the eastern flank of the ReykjanesRidge, which formed by the interaction of IcelandeScotland Over-flow Water (ISOW), flowing as a Deep Western Boundary Current(DWBC), and local topography (Bianchi and McCave, 2000). Duringthe last glaciation, JPC-13 was located a few degrees north of the“Ruddiman IRD belt” (Ruddiman, 1977; Robinson et al., 1995)(Fig. 1). The site is at 3082 m water depth and is bathed today by

e U1308 in the North Atlantic. Surface currents (red lines) and deep currents (blue lines)urning Circulation (AMOC). Light blue shaded area represents the approximate positionthe idealized surface water and dark blue lines the deep-water limbs of AMOC.¼ Western Boundary Undercurrent; CGFZ ¼ Charlie Gibbs Fracture Zone; NAC ¼ Northader is referred to the web version of this article.)

Fig. 2. Stratigraphic overview of Core JPC-13 showing d18O records of C. wuellerstorfi (blue) and N. pachyderma (green). Specimens of N. pachyderma (sin) are too rare for oxygenisotope analysis during the full interglacial conditions of the Holocene and MIS 5e. The chronology of JPC-13 was derived by correlating the benthic d18O of JPC-13 (blue) to MD95-2042 (red) to 80 ka and to LRO4 thereafter (Lisiecki and Raymo, 2005). Also shown for comparison is the Vostok dD record (gray) with Antarctic warm events labeled (A1eA7).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e3886 3877

a mixture of ISOW and Lower Deep-Water (LDW), with the lattercontaining water sourced from the Southern Ocean. ISOW trans-ports approximately 6 sverdrups of water through the CGFZ(Smethie et al., 2007) and, together with Labrador Sea Water andDenmark Strait Overflow Water, constitutes the primary compo-nents of North Atlantic Deep-Water (NADW).

3. Materials and methods

U-channel samples (1.8 cm � 1.9 cm plastic channels cut tolength) were taken from Core JPC-13 from 527 cm to the base of thecore for paleomagnetic studies and XRF scanning. Following theseanalyses, the u-channels were sampled by taking contiguoussamples at 2.5-cm intervals. Each sample was dried, weighed, andwashed over a 63-mm sieve. The sand fraction was dried andweighed to obtain the wt.% greater than 63 mm (wt.% >63 mm). Therelative abundance of ice-rafted detritus was determined by split-ting the >150 mm sediment fraction until approximately 300 lithicgrains remained. IRD abundances are expressed as percent lithicfragments of total grains counted.

Stable isotopic measurements were made on planktonic (Neo-globoquadrina pachyderma sinistral and Globigerina bulloides) andbenthic (Cibicidoides wuellerstorfi) foraminifera. Specimens ofC. wuellerstorfi were selected from the >150 mm size fraction,N. pachyderma (sinistral) from the 150 to 250 mm fraction, andG. bulloides from the 300 to 350 mm size fraction. Foraminifer testswere soaked in w15% H2O2 for 30 min to remove organic matter,and were then rinsed with methanol and sonically cleaned toremove fine-grained particles. The methanol was removed witha syringe, and samples were dried in an oven at 50 �C for 24 h. Theforaminifer calcite was loaded into individual reaction vessels, andeach sample was reacted with three drops of H3PO4 (specificgravity ¼ 1.92) using a ThermoFinnigan MAT Kiel III carbonatepreparation device. Isotope ratios were measured online usinga ThermoFinnigan MAT 252 mass spectrometer. Analytical preci-sion was estimated to be �0.08 for d18O and �0.04& for d13C bymeasuring 8 standards (NBS-19, total ¼ 291) with each set of 38samples. For stable isotopic analysis of bulk carbonate we followedthe method detailed by Hodell and Curtis (2008).

Anhysteretic susceptibility (kARM) divided by volume suscepti-bility (k) is a commonly used grain-size proxy for magnetite in bothmarine and lake sediments (King et al., 1983; Tauxe, 1993). For JPC-13, values of kARM and k were measured on the u-channel samplesusing a 2G Enterprises cryogenic magnetometer and a suscepti-bility track designed for u-channel samples (Thomas et al., 2003).Higher (lower) values of the kARM/k ratio correspond to lower(higher) mean magnetic (magnetite) grain size. The grain-sizecalibration of the kARM/k ratio, following King et al. (1983), indicatesmean magnetite grain sizes below 10 mm for most pelagic sedi-ments (including JPC-13). The kARM/k ratio may, however, beaffected bymagnetic grains in the sortable-silt (10e63 mm) fractioninfluenced by hydrodynamic bottom-current sorting.

High-resolution (2.5-mm) XRF core scanning measurementswere obtained on u-channel subcores at the University of Cam-bridge using an Avaatech XRF core scanner (Richter et al., 2006).The core surface irradiated was 6-mm (cross-core) by 2.5 mm(downcore) with a count time of 40 s. Three scans were performedevery 2.5 mm at 10, 30 and 50 kv and 0.2 mA to obtain both lightand heavy elements. Results are presented as log ratios that providea reliable signal of relative changes in chemical composition(Weltje and Tjallingii, 2008).

4. Chronology

Core JPC-13 extends from the Holocene (MIS 1) to the LastInterglacial (MIS 5e) (Fig. 2). For the interval from 0 to 80 ka, thetime scale (SFCP04) was derived by correlating the benthic d18Orecord of JPC-13 to the benthic d18O record of coreMD95-2042 fromthe Portuguese Margin (Shackleton et al., 2004). The millennial-scale features in benthic d18O are very similar between the two sites(Fig. 2) and also with IODP Site U1308 (Hodell et al., 2008). As notedby Shackleton et al. (2000, 2004), the benthic d18O signal at MD95-2042 (and JPC-13) resembles Antarctic temperature variability (e.g.,warming events A1eA4). The chronology beyond 80 kawas derivedby correlating benthic d18O to the LR04 stack of Lisiecki and Raymo(2005) and is shown for completeness (Fig. 2).

North Atlantic Ash Zones 1 (AZ1) and 2 (AZ2) are identified inJPC-13 providing two additional tie points for correlation to other

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e38863878

North Atlantic marine sediment and Greenland ice cores (Fig. 3). AZ1is centered at 622 cm and is composed of several ashes that arewidely dispersed across the North Atlantic including the Vedde Ash,which is dated at w12.1 ka within the Younger Dryas Chronozone(Lacasse et al., 1995; Zielinski et al., 1997; Thornalley et al., 2010).

Ash Zone 2 (AZ2) occurs at 1151 cm and is dated at 53.26 � 5% ka(Meese et al., 1994), coinciding with the rapid climate transition(cooling) at the end of Greenland Interstadial (GIS) 15 (Austin et al.,2004). The date of AZ2 in our record on the SFCP time scale is 55.7 ka.

A prominent low in relative paleointensity, associated with theLaschamp geomagnetic excursion, occurs at 907.5 cm in core JPC-13. In the North Atlantic, the Laschamp typically lies 20e30 cmbeneath the IRD layer of Heinrich Event 4 (Thouveny et al., 2008).Our placement of Heinrich 4 at 884 cm is consistent with thisobservation. The Laschamp has been correlated to Greenland IS10based on a peak in cosmogenic nuclide abundance and is dated at40.4 � ka (Guillou et al., 2004).

During the last glacial period, sedimentation rates average20 cm kyr�1 but vary from as low as 7 cm kyr�1 to as high as90 cm kyr�1. Given a sampling interval of 2.5 cm, the average

Fig. 3. Ice-rafted detritus proxies from Core JPC-13 including (a) percent total lithics (black)values (upward), and (c) bulk carbonate d18O (gray). Also shown are d18O records of (d) C. wmark the position of Heinrich (H) events, cold (C) periods of McManus et al. (1994), and Ashthe reader is referred to the web version of this article.)

temporal resolution of the record is about 124 years. The SFCP04chronology was used not because it is necessarily the most accurateor current time scale (Skinner, 2008), but rather because bothmarine and polar ice core records can be confidently placed on thistime scale.

5. Results

5.1. Ice-rafted detritus

Ice-rafted detritus layers in JPC-13 are marked by peaks inpercent lithic grains, wt.% coarse fraction, and increases inmagneticgrain size (kARM/k) (Fig. 3). IRD consists dominantly of quartz andvolcanic grains. Coarse lithic peaks occur at each of the Heinrichevents although they lack the large amounts of coarse-graineddetrital carbonate found in the IRD belt to the south (Fig. 1).Increases in fine-grained detrital carbonate are indicated bydecreases in bulk carbonate d18O (Hodell and Curtis, 2008). IRDpeaks are also associated with cold events C19, 20 and 21 identifiedby McManus et al. (1994). Two of the coarse lithic peaks coincide

and volcanics (magenta), (b) magnetic grain size (kARM/k) (green) coarsening for loweruellerstorfi (blue) and (e) N. pachyderma (red) from Core JPC-13. Vertical dashed lineszones (AZ) 1 and 2. (For interpretation of the references to colour in this figure legend,

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e3886 3879

with ash layers (AZ1 and 2) and are dominated by volcanic material(Fig. 3).

5.2. Stable isotopes

The benthic d18O of JPC-13 is marked by a series of millennial-scale oscillations of triangular form beginning with sharp decreasesfollowed by more gradual increases (Fig. 4). The sharp benthic d18Odecreases are labeled A1eA7 assuming these events are synchro-nous with warmings in the Antarctic ice cores as observed byShackleton et al. (2000). Each of the benthic d18O decreases isassociated with an increase in percent IRD and sediment coarsefraction. The benthic d18O variations during MIS 3 are on the orderof 0.3e0.5&. The transition from MIS 5a to 4 is marked by twomillennial-scale decreases (labeled A5 and A6) superimposed upona more gradual trend of increasing d18O. The benthic d18O changefrom MIS 5b to 5a is abrupt as is the transition from MIS 5b to 5cand 5d to 5c with an amplitude of 0.6 and 0.4&, respectively (Figs. 2and 4). The 5ee5d part of the record is discussed by Hodell et al.(2009).

The d18O record of N. pachyderma has a “square-wave” form andshows a variable lag with respect to the benthic d18O signal forevents A1eA7. The d18O of G. bulloides shows more millennial-scalestructure than the d18O of N. pachyderma. Prominent decreases inthe d18O of G. bulloides and N. pachyderma follow each of the IRDpeaks (Fig. 4). This pattern of decreasing planktic d18O followingIRD layers in JPC-13 differs from other North Atlantic records fromthe IRD belt to the south where low planktonic d18O coincides withHeinrich layers (Bond et al., 1993, 1997; Rashid and Boyle, 2007).

6. Discussion

6.1. IRD and surface water hydrography

Core JPC-13 is on the northern margin of the IRD belt (Robinsonet al., 1995). The IRD in the core is composed principally of quartzand basaltic glass with ancillary volcanic and sedimentary rockfragments derived frommultiple sources including Iceland, Europe,North America and Greenland (van Kreveld et al., 2000). Coarse-grained detrital carbonate is lacking but Hodell and Curtis (2008)

Fig. 4. (a) Percent lithics (black) and d18O records of (b) Cibicidoides (blue), (c) left-coiling Nand cold (C) events (McManus et al., 1994), Antarctic warm events (A1eA7), Ash zones (AZ)that coincide with decreases in benthic d18O. (For interpretation of the references to colour

proposed the d18O of bulk carbonate in North Atlantic sedimentsmay be used to recognize sediment layers containing detritalcarbonate, especially near the margins of the IRD belt where fine-fraction detrital carbonate may be transported. Andrews (2000)pointed out that the greatest proportion of ice-rafted sediment isin the clay and silt fractions. It is this fine fraction that is more likelyto be transported greater distances by currents and eddies than iscoarse-grained IRD.

Because the d18O of detrital carbonate from Hudson Straitaverages �5.5& compared to biogenic carbonate of þ4 to þ5&,bulk d18O decreases when detrital carbonate is present (Fig. 5)(Hodell and Curtis, 2008). Each of the peaks in % lithics in Core JPC-13 is associated with a decrease in bulk carbonate d18O, suggestingthey are correlative with Heinrich layers in the IRD belt to the south(Fig. 6). Heinrich Layers 1, 2, 4 and 5 have the lowest bulk d18Oconsistent with a Hudson Strait source for these events (Hemming,2004). We attribute these low bulk d18O values to the occurrence offine-grained detrial carbonate transported from the IRD belt and ofLaurentide origin, although we cannot entirely eliminate additionalcarbonate sources from Europe or older reworked marine sedi-ment. Each of the lithic peaks, including the Heinrich Events, showsa concomitant increase in volcanic material derived from Iceland(Fig. 3) (Lacasse et al., 1998) or possibly East Greenland (Andrewset al., 1999). Volcanic grains have been found to be associatedwith Heinrich events in the IRD belt to the south and precedeslightly the occurrence of detrital carbonate (Bond and Lotti, 1995).

IRD peaks in JPC-13 do not coincide with low planktonic d18O asthey do in the central IRD belt to the south, suggesting they werenot associated with abundant meltwater. Instead, low planktonicd18O immediately follows the IRD peaks (Fig. 4). The pronounceddecreases in d18O of N. pachyderma and G. bulloides following thebenthic d18O decrease and IRD peaks could represent eitherwarmertemperatures and/or reduced surface salinity. The absence of IRDduring the d18O minima argues against iceberg melting unless theice bergs were clean (i.e., devoid of detritus) or melted elsewhereand the low-d18O water was transported to the site (Elliot et al.,1998). The lack of a planktonic d18O decrease associated withHeinrich events in JPC-13 is consistent with results of Cortijo et al.(2000) who found no significant changes in salinity during H4north of 51�N in contrast to cores from the IRD belt to the south

. pachyderma (red) and (d) G. bulloides (black). Also shown are position of Heinrich (H)1 and 2, and Marine Isotope Stages (MIS). Dashed vertical lines designate lithic peaksin this figure legend, the reader is referred to the web version of this article.)

Fig. 5. d18O and d13C of bulk carbonate in Core JPC-13 (black diamonds) and Site U1308 (red circles) compared with values from individual detrital carbonate grains (blue squares)from Heinrich layers (Hodell and Curtis, 2008). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e38863880

between 40 and 50�N. It is also possible that the low planktonicd18O values reported during Heinrich events in the IRD belt havebeen over estimated because of the incomplete removal of detritalcarbonate from the internal chambers of foraminifer tests (Hodelland Curtis, 2008).

Elliot et al. (1998) observed a similar phase relationshipbetween IRD and planktonic d18O in the Irminger Basin where low

Fig. 6. (bottom) Comparison of bulk carbonate (gray) and N. pachyderma d18O (green) focarbonate d18O that mark Heinrich layers. (top) Oxygen isotope records of foraminifera incluand JPC-13 (red). (For interpretation of the references to colour in this figure legend, the re

planktonic d18O occurred a few 100 years after the peak in IRD. Theyinterpreted this lag as reflecting low-salinity water derived fromnorthward propagation of iceberg meltwater produced in the IRDbelt to the south. We instead attribute the low planktonic d18Oevents to the abrupt warmings that immediately followed Heinrichstadials and coincided with the longest, warmest interstadials inGreenland. This interpretation is supported by the warm sea

r IODP Site U1308 and JPC-13. Correlation was derived by matching minima in bulkding C. wuellerstorfi (black) from Core JPC-13 and N. pachyderma from Site U1308 (blue)ader is referred to the web version of this article.)

Fig. 7. Comparison of d18O of N. pachyderma from JPC-13 (blue), G. bulloides from Core MD95-2042 (black), and GRIP ice core (red) from 20 to 80 ka. Both records are on the SFCP04time scale. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e3886 3881

surface temperatures that immediately follow IRD in other coresfrom the Reykjanes Ridge at 59�N (van Kreveld et al., 2000; Jonkerset al., 2010a). Using the SFCP04 time scale, each of the lows in N.pachyderma d18O in Core JPC-13 correlates with a similar magnitudedecrease in planktonic d18O in IberianMargin CoreMD95-2042 andwith long Greenland interstadials (Fig. 7). The low planktonic d18Oat JPC-13 also correlates to those at Site U1308 (redrill of DSDP Site609) in the IRD belt to the south (Fig. 6) (Hodell and Curtis, 2008).

The d18O of N. pachyderma following Heinrich events decreasesby w1& during MIS3 (Fig. 8). If attributed all to temperature, thiswould suggest w4 �C of warming just below the base of the

Fig. 8. Percent lithics (black open circles), wt.% >63 mm fraction (gray), and d18O records of Ccore (black line), and pCO2 composite from Antarctic ice cores (open circles from Byrd, filledlithic peaks including Heinrich events, C-events and Ash Zones 1 and 2. All records on SFCcolour in this figure legend, the reader is referred to the web version of this article.)

seasonal mixed layer where N. pachyderma (sin) calcifies (Jonkerset al., 2010a, b). Warm winter temperatures are also supported byforaminiferal transfer functions (van Kreveld et al., 2000) and d15Nof gas from Greenland ice cores that typically indicate a rapidwarming of w12 �C at the start of the long interstadials followingHeinrich events (Huber et al., 2006).

In cores at 59�N on the Reykjanes Ridge, Jonkers et al. (2010a)found that warming of surface water began earlier during thetime of IRD deposition and preceded the main warming associatedwith the pronounced interstadial following the Heinrich event (seealso Moros et al., 2002). They attributed the warm temperatures

ibicidoides (blue crosses), left-coiling N. pachyderma (red crosses), dD of the Vostok icecircles from Taylor Dome, filled squares from Vostok). Dashed vertical lines designate

P04 time scale expect CO2 which is on EDC-3. (For interpretation of the references to

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e38863882

during IRD events at w59�N to the presence of a subsurface warmlayer that seasonally shoaled during summer. We do not observedecreasing d18O of N. pachyderma or G. bulloides during IRD depo-sition at 53�S, but our d18O results strongly support the tempera-ture maxima that occurred after IRD deposition ceased (vanKreveld et al., 2000; Jonkers et al., 2010a). Mg/Ca and faunalabundance counts of foraminifera are needed in Core JPC-13 todetermine if seasonal warming during IRD deposition may havepreceded the main planktonic d18O decrease.

6.2. Phasing of isotopic and IRD proxies

The relative phasing of variables can be observed in the depthdomain and is independent of the age model applied (Fig. 8). Wefind a similar phasing in surface and deep-water d18O as reportedby Shackleton et al. (2000); that is, benthic d18O leads planktonicd18O (Figs. 4 and 8). Each of the peaks in % lithics coincides witha decrease in benthic d18O followed by a decrease in the d18O ofNeogloboquadrina pachyderma and G. bulloides (Figs. 6 and 8). Thisphasing supports previous results indicating the coldest stadials inGreenland (e.g., Heinrich stadials) occurred during times ofwarming in Antarctica with the onset of GIS events 8, 12, 14, 16/17,19, 20, and 21 by 1.5e3 kyr (Blunier and Brook, 2001; Capron et al.,2009).

Although determining the phase relationships in a single coreavoids some of the problems associated with time scale inaccuracy,it is not without other complications related to changes in sedi-mentation rate, sampling resolution, and bioturbation that canaffect the nature and timing of paleoenvironmental signals. Forexample, lower sedimentation rates will attenuate the signal andobscure phasing among variables. Mean sedimentation rates arehigh in Core JPC-13, averaging 20 cm kyr�1, but interval rates varyfrom as low as 7 cm kyr�1 to as high as 90 cm kyr�1.

The N. pachyderma d18O records the warmest interstadials in theGreenland ice core (GIS8, 12, 14, 16e17) immediately following theHeinrich events (Bond et al., 1993), but the shorter, interveninginterstadials are less well defined. This pattern may reflect thegreater attenuation of signal for the shorter-term Dans-gaardeOeschger (DeO) events than for the longer-duration inter-stadials. For example, GIS 8, 12 and 14 and 16e17 were nearly twiceas long as the other interstadials.

Anderson (2001) estimated a 70% reduction in amplitude ofshort-term DeO events for sedimentation rates of 10 cm kyr�1. Forlonger-duration events, sedimentation rates of 10 cm kyr�1

preserve 50% of the original signal, whereas 95% of the originalsignal is captured at 50 cm kyr�1 (Anderson, 2001).

During times of rapid change, such as Heinrich stadials followedby abrupt warming, phase shifts can be induced for species withopposite abundance patterns (Hutson, 1980; Löwenmark andGrootes, 2004; Löwemark et al., 2008). For example, the d18Odecrease following Heinrich events at JPC-13 occurs earlier inG. bulloides than N. pachyderma (sin) (Fig. 4). N. pachyderma (sin)was more abundant during the Heinrich stadials in the subpolarNorth Atlantic, whereas G. bulloides was more prevalent duringGreenland interstadials (Van Kreveld et al., 2000). Bioturbationwilltend to move particles from levels of higher to lower concentrationthereby resulting in a mixing down of isotopically lighter speci-mens of G. bulloides into the underlying IRD layer and mixing up ofstadial N. pachyderma specimens into the interstadial. Bioturbationcoupledwith changing species abundance can explain the apparentlead in G. bulloides d18O relative to N. pachyderma (sin) (Fig. 4).

It is difficult to evaluate how changes in the relative abundanceof benthic foraminifer taxa might influence phasing because theirabundance and species composition are not only affected byphysio-chemical properties of water masses but also by organic

carbon rain from the surface. It is unlikely the benthic lead overplanktonic d18O is an artefact of bioturbation and changing speciesabundance in MD95-2042 or JPC-13 because the phasing is inde-pendently supported bymethane synchronization of polar ice cores(Blunier and Brook, 2001; EPICA Community Members, 2006).

Establishment of the phase relationships for the last glacialperiod in Core JPC-13 is important because it paves the way forapplying similar methods to study phase relationships in olderglacial periods of the Pleistocene at IODP Site U1304 drilled at thesame location. The Site U1304 record extends to the top of theOlduvai Subchronozone (1.77 Ma) (Expedition 303 Scientists,2006).

6.3. A pervasive Antarctic-type signal in the North Atlantic?

Pisias et al. (2010) demonstrated two leading modes of low-frequency global climate variability during MIS 3: a Greenland or“northern” mode and an Antarctic or “southern” mode (see alsoClark et al., 2007). To the extent that benthic d18O reflects anAntarctic signal, IRD peaks in the North Atlantic also displaya distinct Antarctic-type pattern. We suggest that the d18O ofN. pachyderma also shows an Antarctic-style pattern, albeit witha time delay with only the largest and longest interstadialsfollowing Antarctic warming events clearly expressed in theplanktonic d18O record of the subpolar North Atlantic (Fig. 8).

Barker and Knorr (2007) similarly suggested that Antarctic-styleclimate variability is a globally pervasive pattern that is evenembedded in Greenland ice cores. Increases in atmospheric CO2accompanied the millennial-scale warming events in Antarcticaduring MIS 3, thereby providing a mechanism for global propaga-tion of the Antarctic climate signal (Fig. 8) (Indermühle et al., 2000;Ahn and Brook, 2008).

The Antarctic signal is prevalent in subpolar surface waters butis delayed because iceberg melting and sea-ice expansion duringHeinrich events kept the subpolar North Atlantic cold despitewarming elsewhere (Moros et al., 2002). Because the low-salinitysurface layer was underlain by a warm subsurface water (Mignotet al., 2007), abrupt warming ensued once the rate of icebergcalving diminished and surface stratification broke down.Schmittner et al. (2002) used an Earth System Climate Model toshow that reduced iceberg calving following Heinrich events canact as a trigger for rapid warming within several 100 years. Jonkerset al. (2010a) similarly suggested that the abrupt warmingfollowing Heinrich events was related to the rapid emptying ofa subsurface heat reservoir that restarted overturning circulation.

6.4. Northward heat transport and deep-water circulation

A change in the strength of AMOC is commonly invoked toexplain anti-phasing of millennial-scale climate variability betweenGreenland and Antarctica (Broecker, 1998; Alley, 2007) althoughsea-ice and tropical processes may also play important roles(Clement and Peterson, 2008). AMOC was reduced during Heinrichevents as lowered sea surface salinity decreased overturning andpermitted the southern expansion of sea ice in the North Atlantic.This process is expressed in benthic d13C variations in Core MD95-2042 that show lower values during stadials and higher valuesduring interstadials (Fig. 9) (Skinner et al., 2007; Martrat et al.,2007). The IRD peaks (Heinrich events) in JPC-13 coincide withbenthic d13C minima in MD95-2042 indicating a link betweensurface-deep North Atlantic hydrography (Fig. 9). Benthic d13C inCore JPC-13 also decreases across H1, H2, H4 and H5 although thepattern is not as clear as MD95-2042 (Fig. 10).

Several lines of evidence in JPC-13 suggest that ISOW wasstrongerduring thewarmer, longerGreenland interstadials (e.g., GIS

Fig. 9. Benthic d13C of Core MD95-2042 Shackleton et al. (2004) (blue) compared with percent lithics (red) and d18O of N. pachyderma (black) from Core JPC-13. The IRD peaks inJPC-13 are generally associated with lows in benthic d13C (note scale decreases in upwards direction). (For interpretation of the references to colour in this figure legend, the readeris referred to the web version of this article.)

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e3886 3883

8, 12, and 14) that immediately followed Heinrich events. Abruptwarming in Greenland following Heinrich events coincided withincreases in benthic d13C in Core MD95-2042 and decreases insediment grain size in Core JPC-13 (Fig. 11). During each of the longGreenland interstadials, sediments are dominated by silt and clayand the wt.% coarse fraction is low in Core JPC-13. The sand fraction(>63 mm) on the southern Gardar Drift consists of foraminifera andIRD, and is strongly diluted by silt and clay transported by ISOW(Bianchi and McCave, 2000). We suggest the decreases in grain sizeduringMIS 3 are the result of increased transport offine sediment byISOWassociated with the longest, warmest Greenland interstadials(GIS 8, 12,14, 16e17). Magnetic grain size (kARM/k) also decreases atthese times suggesting increased transport of fine-grained magne-tite by ISOW to the southern end of the Gardar Drift (Fig. 3). We do

Fig. 10. Benthic d13C (red; 3-pt running average) and % lithics (blue) in Core JPC-13. (For inteweb version of this article.)

not believe that kARM/k reflects current speed as grains contributingthis parameter are probably too small to be hydrodynamically sor-ted. The parameter probably primarily denotes changes in thecontribution of fine grained detritus from northern (Icelandic)sources (see Kissel et al., 2009).

Further support for changes in ISOW is provided by K/Ti ratiosthat reflect the relative contribution of acid (felsic) and basic(mafic) sources of terrigenous sediment (Prins et al., 2001; Richteret al., 2006; Ballini et al., 2006; Grützner and Higgins, in press). LowK/Ti in Core JPC-13 occurred during the Holocene and MIS 5 whenISOW was strong and transported fine-grained sediment southfrom the Iceland basaltic province (Fig. 11), thereby building theGardar Drift. K/Ti in Core JPC-13 was also relatively low during eachof the long, warm interstadials of MIS 3 when increasing benthic

rpretation of the references to colour in this figure legend, the reader is referred to the

Fig. 11. (a) % lithics and (b) Sr/Zr (and IRD proxy) and (c) wt.% >63 mm sediment coarse fraction in Core JPC-13 (d) benthic d13C in Core MD95-2042; (e) Greenland ice core d18O; and(f) logK/Ti in Core JPC-13. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e38863884

d13C and decreasing grain size also support increased convection inthe Nordic Seas (Fig. 11).

Results from Core JPC-13 support previous findings for increasedheat transport to the high latitudes, northward penetration of theNorth Atlantic Current, and increased meridional overturningduring the warmest, longest Greenland interstadials of MIS 3(Piotrowski et al., 2008; Kissel et al., 2008). Strong northward heattransport and warming of the subpolar North Atlantic andGreenland following Antarctic warm events may have been relatedto the abrupt resumption of AMOC following Heinrich events(Schmittner et al., 2002; Knorr and Lohmann, 2007; Liu et al.,2009).

6.5. Benthic d18O, ice volume, and deep hydrographic change

Results from Core JPC-13 confirm that decreases in benthicd18O coincided with peaks in IRD during the last glaciation on thesouthern Gardar Drift, and were correlated with warmings eventsin Antarctica (A1eA7). The Antarctic-like, millennial-scale varia-tions in benthic d18O appear to be a robust feature of NorthAtlantic records at water depths between 3 and 4 km (Fig. 2).Although independent studies indicate that climate variability inthe high-latitude North Atlantic and Antarctica were tightlycoupled during the last glaciation (e.g., EPICA CommunityMembers, 2006), it is not entirely clear why the benthic d18Osignal in the North Atlantic has the same character and phasing asAntarctic ice cores.

Changes in the d18O of benthic foraminifera reflect a combina-tion of temperature, global ice volume (sea level), and localhydrographic effects (Waelbroeck et al., 2002; Skinner et al., 2003).

There is considerable uncertainty, however, over the relativeimportance of these processes. Shackleton et al. (2000) originallyinterpreted the millennial changes in benthic d18O during MIS 3 onthe Iberian Margin as indicating significant variations in icevolume.

The benthic d18O variations at JPC-13 are on the order of0.3e0.5& for A1eA4, which would permit up to 30e50 m of sea-level variation if all the change was attributed to ice volume. Sea-level rise during the warm events in Antarctica is supported by theRed Sea d18O record, which indicates rather large sea-level varia-tions that are similar to Antarctic-style climate variations (Siddallet al., 2003, 2008; Rohling et al., 2004, 2008) proposed at leastfour millennial-scale sea-level fluctuations of 20e30 m magnitudeduring MIS 3 with rises occurring during warming events inAntarctica and cold stadial conditions in Greenland (see alsoGonzalez and Dupont, 2009). Sea-level rise during the Antarcticwarm events may have been related to both dynamic and stericeffects that, in turn, may have triggered a dynamic response of theLaurentide Ice Sheet (e.g., Flückiger et al., 2006). However, otherstudies have suggested that sea-level highstands coincided withGreenland interstadials and post-dated Antarctic warming (Arzet al., 2007; Sierro et al., 2009).

Skinner et al. (2003, 2007) demonstrated that only part of thebenthic d18O signal in MIS 3 could be related to changes in conti-nental ice volume with the remainder attributed to temperatureand hydrographic change. Skinner and Elderfield (2007) foundabrupt warming of deep-water during the middle of Heinrich sta-dials 4 and 5 indicating important changes in deep-sea temperaturedistribution. A subsurface heat reservoir at depth in the NorthAtlantic may have contributed to the rapid warming that followed

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e3886 3885

Heinrich events once breakdown of surface stratification occurred(Schmittner et al., 2002; Shaffer et al., 2004; Mignot et al., 2007).

This study demonstrates that each of the benthic d18O decreasesin JPC-13 coincides with a peak in IRD, suggesting a possible linkwith ice bergs and/or sea ice (Fig. 4). Van Kreveld et al. (2000)proposed that decreasing benthic d18O in the Irminger Sea, whichoccurred just prior to the abrupt warmings in Greenland, may berelated to a brine water pulse induced by seasonal freezing of largeparts of the northwestern Atlantic during Heinrich events. Hillaire-Marcel and de Vernal (2008) similarly proposed that light isotopicexcursions in N. pachyderma associated with Heinrich events couldbe related to episodes of enhanced sea-ice formation. The mixing ofcold, low-d18O brine water and poorly ventilated deep-water fromthe Southern Ocean during Heinrich events might explain theoccurrence of a cold, low-d18O and low-d13C water mass duringHeinrich stadials (Skinner et al., 2007). Explaining why the benthicd18O signal so closely follows an Antarctic-style pattern remains animportant problem for understanding mechanisms of interhemi-spheric climate change.

7. Conclusions

Results from Core JPC-13 demonstrate that decreases in benthicd18O coincided with peaks in IRD during the last glaciation on thesouthern Gardar Drift. The IRD events are correlated with warmingevents in Antarctica (A1eA7), supporting a coupling of climatevariability between the high-latitude North Atlantic and Antarcticaduring the last glaciation. The Antarctic warmings correspond tothe coldest stadials in Greenland, associated with Heinrich eventsat the end of Bond cycles. The IRD pulses are followed shortlythereafter by abrupt decreases in planktonic d18O indicatingwarming of SST that is correlated with the most pronouncedinterstadials in Greenland. These changes coincided with anincreased strength of AMOC and associated heat transport to thesubpolar North Atlantic. The observed phasing in Core JPC-13 is thesame as that inferred from marine sediment cores on the Portu-guese Margin (Shackleton et al., 2000) and polar ice coressynchronized using variations in atmospheric methane (Blunierand Brook, 2001). The method applied here can be extended toIODP Site U1304, at the same location as Core JPC-13, to study phaserelationships during older glacial periods of the Pleistocene.

Acknowledgments

We thank Gianna Browne for technical assistance and JimWright and Greg Mountain for collecting Core KN166-14-13JPC andproviding samples. We thank Peter Clark and an anonymousreferee whose thoughtful reviews significantly improved thequality of the manuscript. This research was supported by NERCgrant NE/H009930/1.

References

Ahn, J., Brook, E.J., 2008. Atmospheric CO2 and climate on millennial time scalesduring the last glacial period. Science 322, 83e85.

Alley, R.B., 2007. Wally was right: predictive ability of the North Atlantic “conveyorbelt” hypothesis for abrupt climate change. Annu. Rev. Earth Planet. Sci. 35,241e272.

Anderson, D.A., 2001. Attenuation of millennial-scale events by bioturbation inmarine sediments. Paleoceanography 16, 352e357.

Andrews, J.T., 2000. Icebergs and iceberg rafted detritus (IRD) in the North Atlantic:facts and assumptions. Oceanography 13, 100e108.

Andrews, J.T., Barber, D.C., Jennings, A.E., 1999. Errors in generating time-series andin dating events at late Quaternary millennial (radiocarbon) time-scales:examples from Baffin Bay, NW Labrador Sea, and East Greenland. Mechanismsof global climate change at millennial time scales. Geophys. Monogr. 112,23e34.

Arz, H.W., Lamy, F., Ganopolski, A., Nowaczyk, N., Patzold, J., 2007. Dominantnorthern hemisphere climate control over millennial-scale glacial sea-levelvariability. Quaternary Sci. Rev. 26, 312e321.

Austin, W.E.N., Wilson, L.J., Hunt, J.B., 2004. The age and chronostratigraphicalsignificance of north Atlantic ash zone II. J. Quat. Sci. 19 (2), 137e146.

Ballini, M., Kissel, C., Colin, C., Richter, T., 2006. Deep-water mass source anddynamic associated with rapid climatic variations during the last glacial stage inthe North Atlantic: a multiproxy investigation of the detrital fraction of deep-sea sediments. Geochem. Geophys. Geosyst. 7 (Q02N01). doi:10.1029/2005GC001070.

Barker, S., Knorr, G., 2007. Antarctic climate signature in the Greenland ice corerecord. Proc. Nat. Acad. Sci. U.S.A. 104, 17278e17282.

Bianchi, G.G., McCave, I.N., 2000. Hydrography and sedimentation under the deepwestern boundary current on Björn and Gardar Drifts, Iceland Basin. Mar. Geol.165, 137e169.

Blunier, T., Brook, E.J., 2001. Timing of millennial-scale climate change in Antarcticaand Greenland during the last glacial period. Science 291, 109e112.

Blunier, T., Chappellaz, J., Schwander, J., Dallenbach, A., Stauffer, B., Stocker, T.F.,Raynaud, D., Jouzel, J., Clausen, H.B., Hammer, C.U., Johnsen, S.J., 1998. Asyn-chrony of Antarctic and Greenland climate change during the last glacial period.Nature 394, 739e743.

Bond, G.C., Lotti, R., 1995. Iceberg discharges into the North Atlantic on millennialtime scales during the last glaciation. Science 267, 1005e1010.

Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P., Priore, P.,Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in NorthAtlantic Holocene and glacial climates. Science 278, 1257e1266.

Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L., Jouzel, J., Bonani, G.,1993. Correlations between climate records from North Atlantic sediments andGreenland ice. Nature 365, 143e147.

Broecker, W.S., 1998. Paleocean circulation during the last deglaciation: a bipolarseesaw? Paleoceanography 13 (2), 119e121.

Capron, E., Landais, A., Lemieux-Dudon, B., Schilt, A., Masson-Demotte, V.,Buiron, D., Chappelaz, J., Dahl-Jensen, D., Johnsen, S., Leuenberger, M.,Loulergue, L., Oerter, H., Chappelas, J., 2009. Synchronising EDML and North-GRIP ice cores using d18O of atmospheric oxygen (d18Oatm) and CH4measurements over MIS5 (80e123 kyr). Quaternary Sci. Rev.. doi:10.1016/j.quascirev.2009.07.014.

Charles, C.D., Lynch-Stieglitz, J., Ninnemann, U.S., Fairbanks, R.G., 1996. Climateconnections between the hemisphere revealed by deep sea sediment core/icecore correlations. Earth Planet. Sci. Lett. 142, 19e27.

Clark, P.U., Hostetler, S.W., Pisias, N.G., Schmittner, A., Meissner, K.J., 2007. Mecha-nisms for a w7-kyr climate and sea-level oscillation during marine isotopestage 3. In: Schmittner, A., Chiang, J., Hemming, S. (Eds.), Ocean Circulation:Mechanisms and Impacts. Geophysical Monograph, vol. 173. AmericanGeophysical Union, Washington, D.C, pp. 209e246.

Clement, A.C., Peterson, L.C., 2008. Mechanisms of abrupt climate change of the lastglacial period. Rev. Geophys. 46 (RG4002), 1e39.

Cortijo, E., Labeyrie, L., Vidal, L., Vautravers, M., Chapman, M., Duplessy, J.-C.,Elliot, M., Arnold, M., Turon, J.-L., Auffret, G., 2000. Rapid climatic variability ofthe North Atlantic Ocean and global climate: a focus of the IMAGES program.Quaternary Sci. Rev. 19, 227e241.

Crowley, T.J., 1992. North Atlantic deep water cools the southern hemisphere.Paleoceanography 7, 489e497.

Elliot, M., Labeyrie, L., Bond, G., Cortijo, E., Turon, J.-L., Tisnerat, N., Duplessy, J.-C.,1998. Millennial-scale iceberg discharges in the Irminger Basin during the lastglacial period: relationship with the Heinrich events and environmentalsettings. Paleoceanography 13 (5), 433e446.

EPICA Community Members, 2006. One-to-one coupling of glacial climate vari-ability in Greenland and Antarctica. Nature 444, 195e198.

Expedition 303 Scientists, 2006. SiteU1304. In: Channell, J.E.T., Kanamatsu, T., Sato, T.,Stein, R., Alvarez Zarikian, C.A.,Malone,M.J., Expedition 303/306 Scientists (Eds.),Proc. IODP, 303/306. Integrated Ocean Drilling Program Management Interna-tional, Inc., College Station TX. doi:10.2204/iodp.proc.303306.104.2006.

Flückiger, J., Knutti, R., White, J.W.C., 2006. Oceanic processes as potential triggerand amplifying mechanisms for Heinrich events. Paleoceanography 21(PA2014). doi:10.1029/2005PA001204.

Guillou, H., Singer, B.S., Laj, C., Kissel, C., Scaillet, S., Jicha, B.R., 2004. On the age ofthe Laschamp geomagnetic excursion. Earth Planet. Sci. Lett. 227, 331e343.

Gonzalez, C., Dupont, L.M., 2009. Tropical salt marsh succession as sea-level indi-cator during Heinrich events. Quaternary Sci. Rev. 28, 939e946.

Grützner, J., Higgins, S.M. Threshold behavior of millennial scale variability in deepwater hydrography inferred from a 1.1 Ma long record of sediment provenanceat the southern Gardar Drift. Paleoceanography, in press.

Heinrich, H., 1988. Origin and consequences of cyclic ice rafting in the northeastAtlantic Ocean during the past 130,000 years. Quaternary Res. 29, 142e152.

Hemming, S.R., 2004. Heinrich events: massive late Pleistocene detritus layers ofthe North Atlantic and their global climate imprint. Rev. Geophys. 42 (RG1005).doi:10.1029/2003RG000128.

Hillaire-Marcel, C., de Vernal, A., 2008. Stable isotope clue to episodic sea iceformation in the glacial North Atlantic. Earth Planet. Sci. Lett. 268, 143e150.

Hodell, D.A., Curtis, J.H., 2008. Oxygen and carbon isotopes of detrital carbonate inNorth Atlantic Heinrich Events. Mar. Geol. 256, 30e35.

Hodell, D.A., Minth, E.K., Curtis, J.H., Hall, I.R., Channell, J.E.T., Xuan, C., 2009. Surfaceand deep water hydrography on Gardar drift (Iceland Basin) during the LastInterglacial period. Earth Planet. Sci. Lett. 288, 10e19.

D.A. Hodell et al. / Quaternary Science Reviews 29 (2010) 3875e38863886

Hodell, D.A., Channell, J.E.T., Curtis, J., Romero, O., Röhl, U., 2008. Onset of “HudsonStrait” Heinrich events in the eastern north Atlantic at the end of the middlePleistocene transition (w640 ka)? Paleoceanography 23. doi:10.1029/2008PA001591.

Huber, C., Leuenberger, M., Spahni, R., FlückigerSchwander, J., Stocker, T.F.,Johnsen, S., Landais, A., Jouzel, J., 2006. Isotope calibrated Greenland tempera-ture record over Marine isotope stage 3 and its relation to CH4. Earth Planet. Sci.Lett. 243, 504e519.

Hutson, W.M., 1980. Bioturbation of deep-sea sediments: oxygen isotopes andstratigraphic uncertainty. Geology 8, 127e130.

Indermühle, A., Monnin, E., Stauffer, B., Stocker, T.F., Wahlen, M., 2000. AtmosphericCO2 concentration from 60 to 20 kyr BP from the Taylor Dome ice core,Antarctica. Geophys. Res. Lett. 27 (5), 735e738.

Jonkers, L., Moros, M., Prins,M.A., Dokken, T., AnderssonDahl, C, Dijkstra, N., Perner, K.,Brummer, G.-J.A., 2010a. A reconstruction of sea surface warming in the northernNorth Atlantic during MIS 3 ice-rafting events. Quat. Sci. Rev. 29, 1791e1800.

Jonkers, L., Brummer, G.-J.A., Peters, F.J.C., van Aken, H.M.I., DeJong, M.F., 2010b.Seasonal stratification, shell flux and oxygen isotope dynamics of left-coilingN. pachyderma and T. quinqueloba in the western subpolar North Atlantic.Paleoceanography 25 (PA2204). doi:10.1029/2009PA001849.

King, J.W., Banerjee, S.J., Marvin, J., 1983. A new rock-magnetic approach to selectingsediments for geomagnetic paleointensity studies: application to paleointensityfor the last 4000 years. J. Geophys. Res. 88, 5911e5921.

Kissel, C., Laj,C.,Mulder, T.,Wandres, C.,Cremer,M., 2009. Themagnetic fraction:a tracerofdeepwatercirculation in theNorthAtlantic. EarthPlanet. Sci. Lett. 288, 444e454.

Kissel, C., Laj, C., Piotrowski, A.M., Goldstein, S.L., Hemming, S.R., 2008. Millennial-scale propagation of Atlantic deep waters to the glacial Southern ocean. Pale-oceanography 23 (PA2102). doi:10.1029/2008PA001624.

Knorr, G., Lohmann, G., 2007. Rapid transitions in the Atlantic thermohalinecirculation triggered by global warming and meltwater during the last degla-ciation. Geochem. Geophys. Geosyst. 8 (12). doi:10.1029/2007GC001604.

Knutti, R., Flückiger, J., Stocker, T.F., Timmermann, A., 2004. Strong hemisphericcoupling of glacial climate through freshwater discharge and ocean circulation.Nature 420, 851e856.

Lacasse, C., Carey, S., Sigurdsson, H., 1998. Volcanogenic sedimentation in theIceland Basin: influence of subaerial and subglacial eruptions. J. Volcanol.Geoth. Res. 83, 47e73.

Lacasse, C., Sigurdsson, H., Johannesson, H., Paterne, M., Carey, S., 1995. Source ofash zone 1 in the north Atlantic. Bull. Volcanol. 57, 18e32.

Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene-Pleistocene stack of 57 globallydistributed benthic d18O records. Paleoceanography 20 (PA1003). doi:10.1029/2004PA001071.

Liu, Z., Otto-Bleisner, B.L., He, F., Brady, E.C., Tomas, R., Clark, P.U., Carlson, A.E.,Lynch-Stieglitz, J., Curry, W., Brook, E., Efickson, D., Jacob, R., Kutzbach, J.,Cheng, J., 2009. Transient simulation of last deglaciation with a newmechanismfor Bølling-Allerød warming. Science 325, 310e314.

Löwemark, L., Grootes, P.M., 2004. Large age differences between planktic fora-minifers caused by abundance variations and Zoophycos bioturbation. Paleo-ceanography 19, PA2001. doi:10.1029/2003PA000949.

Löwemark, L., Konstantinou, K.I., Steinke, S., 2008. Bias in foraminiferal multispeciesreconstructions of paleohydrographic conditions caused by foraminiferalabundance variations and bioturbational mixing: a model approach. Mar. Geol.256, 101e106.

Martrat, B., Grimalt, J.O., Shackleton, N.J., de Abreu, L., Hutterli, M.A., Stocker, T.F.,2007. Four climate cycles of recurring deep and surface water destabilizationson the Iberian Margin. Science 317, 502e507.

McManus, J.F., Bond, G.C., Broecker, W.S., Johnsen, S., Labeyrie, L., Higgins, S., 1994.High-resolution climate records from the North Atlantic during the Last Inter-glacial. Nature 371, 326e329.

Meese, D.A., Alley, R.B., Gow, A.J., Grootes, P., Mayewski, P.A., Ram, M., Taylor, K.C.,Waddington, E.D., Zielinski, G., 1994. Preliminary Depth-age Scale of the GISP2Ice Core. CRREL, Special Report, 94-1, 66 pp.

Mignot, J., Ganopolski, A., Levermann, A., 2007. Atlantic subsurface temperatures:response to a shutdown of the overturning circulation and consequences for itsrecovery. J. Clim. 20 (9), 4884e4898.

Mix, A.C., Ruddiman, W.F., McIntyre, A., 1986. Late Quaternary paleoceanography ofthe tropical Atlantic, 1: spatial variability of annual mean sea-surface temper-atures, 0e20,000 years B.P. Paleoceanography 1, 43e66.

Moros, M., Kuijpers, A., Snowball, I., Lassen, S., Backstrom, D., Gingele, F.,McManus, J., 2002. Were glacial iceberg surges in the North Atlantic triggeredby climatic warming? Mar. Geol. 192, 393e417.

Piotrowski, A.M., Goldstein, S.L., Hemming, S.R., Fairbanks, R.G., Zylberberg, D.R.,2008. Oscillating glacial northern and southern deep water formation fromcombined neodymiumand carbon isotopes. Earth Planet. Sci. Lett. 272, 394e405.

Pisias, N.G., Clark, P.U., Brook, E.J., 2010. Modes of climate variability during MarineIsotope Stage 3 (60e26 ka). J. Clim. 23, 1581e1588.

Prins, M.A., Troelstra, S.R., Kruk, R.W., van der Borg, K., de Jong, A.F.M., Weltje, G.J.,2001. The late Quaternary sedimentary record of Reykjanes Ridge, northAtlantic. Radiocarbon 43, 939e947.

Rashid, H., Boyle, E.A., 2007. Mixed-layer deepening during Heinrich events:a multi-planktonic foraminiferal d18O approach. Science 320, 439e441.

Richter, T.O., van der Gaast, S., Koster, B., Vaars, A., Gieles, R., de Stigter, H.C.,DeHaas, H., van Weering, T.C.E., 2006. The Avaatech XRF Core Scanner: Tech-nical Description and Applications to NE Atlantic Sediments. In: GeologicalSociety, London, Special Publications, vol. 267, pp. 39e50.

Robinson, S.G., Maslin, M.A., McCave, I.N., 1995. Magnetic susceptibility variations inupper Pleistocene deep-sea sediments of the NE Atlantic: implications for icerafting and paleocirculation at the last glacial maximum. Paleoceanography 10,221e250.

Rohling, E.J., Grant, K., Hemleben, C., Kucera, M., Roberts, A.P., Schmeltzer, I.,Schulz, H., Siccha, M., Siddall, M., Trommer, G., 2008. New constraints on thetiming of sea level fluctuations during early to middle marine isotope stage 3.Paleoceanography 23 (PA3219). doi:10.1029/2008PA001617.

Rohling, E.J., Marsh, R., Wells, N.C., Siddall, M., Edwards, N.R., 2004. Similar melt-water contributions to glacial sea level changes from Antarctic and northern icesheets. Nature 430, 1016e1021.

Ruddiman, W.F., 1977. Late Quaternary deposition of ice-rafted sand in the subpolarNorth Atlantic (lat 40� to 65�N). Geol. Soc. Am. Bull. 88, 1813e1827.

Schmittner, A., Yoshimori, M., Weaver, A.J., 2002. Instability of glacial climate ina model of the ocean-atmosphere-cryosphere system. Science 295,1489e1493.

Shackleton, N.J., Fairbanks, R.G., Chiu, T.-C., Parrenin, F., 2004. Absolute calibrationof the Greenland time scale: implications for Antarctic time scales and for D14C.Quaternary Sci. Rev. 23, 1513e1522.

Shackleton, N.J., Hall, M.A., Vincent, E., 2000. Phase relationships betweenmillennial-scale events 64,000e24,000 years ago. Paleoceanography 15 (6),565e569.

Shaffer, G., Olsen, S.M., Bjerrum, C.J., 2004. Ocean subsurface warming as a mech-anism for coupling DansgaardeOeschger climate cycles and ice-rafting events.Geophys. Res. Lett. 31 (L24202). doi:10.1029/2004GL020968.

Siddall, M., Rohling, E.J., Thompson, W.G., Waelbroeck, C., 2008. MIS 3 Sea-levelfluctuations: data synthesis and new outlook. Rev. Geophys. 46 (RG4003).doi:10.1029/2007RG000226.

Siddall, M., Rohling, E.J., Almogi-Labin, A., Hemleben, Ch., Meischner, D.,Schmelzer, I., Smeed, D.A., 2003. Sea-level fluctuations during the last glacialcycle. Nature 423, 853e858.

Sierro, F.J., Andersen, N., Bassetti, M.A., Berne, S., Canals, M., Curtis, J.H.,Dennielou, B., Flores, J.A., Figola, J., Gonzalez-Mora, B., Grimalt, J.O., Hodell, D.A.,Jouet, G., Perez-Folgado, M., Schneider, R., 2009. Quaternary Sci. Rev. 28, 1e15.

Skinner, L.C., 2008. Revisiting the absolute calibration of the Greenland ice-coreage-scales. Clim. Past Discuss. 4, 791e807.

Skinner, L.C., Elderfield, H., 2007. Rapid fluctuations in the deep North Atlantic heatbudget during the last glacial period. Paleoceanography 22 (PA1205).doi:10.1029/2006PA001338.

Skinner, L.C., Elderfield, H., Hall, M., 2007. Phasing of millennial events andNortheast Atlantic deep-water temperature change since 50 ka BP. In: OceanCirculation: Mechanisms and Impacts e Past and Future Changes of MeridonialOverturning. AGU, Washington D.C., pp. 197e208.

Skinner, L.C., Shackleton, N.J., Elderfield, H., 2003. Millennial-scale variability ofdeep-water temperature and d18Odw indicating deep-water source variations inthe Northeast Atlantic, 0e34 cal. ka BP. Geochem. Geophys. Geosyst. 4 (12),1098. doi:10.1029/2003GC000585.

Smethie, W.M., LeBel, D.A., Fine, R.A., Rhein, M., Kieke, D., 2007. Strength andvariability of the deep limb of the North Atlantic meridional overturningcirculation from chlorofluorocarbon inventories. In: Schmittner, A., Chiang, J.,Hemming, S. (Eds.), Ocean Circulation: Mechanisms and Impacts. GeophysicalMonograph Series, vol. 173, pp. 119e130.

Steig, E.J., Alley, R.B., 2002. Phase relationships between Antarctic and Greenlandclimate records. Ann. Glaciol. 35, 451e456.

Stocker, T.F., 1998. The seesaw effect. Science 282, 61e62.Stocker, T.F., Johnsen, S.J., 2003. A minimum model for the bipolar seesaw. Paleo-

ceanography 18, 1087. doi:10.1029/2003PA000920.Tauxe, L., 1993. Sedimentary records of relative paleointensity of the geomagnetic

field: theory and practice. Rev. Geophys. 31, 319e354.Thomas, R., Guyodo, Y., Channell, J.E.T., 2003. U-channel track for susceptibility

measurements. Geochem. Geophys. Geosyst., 1050. doi:10.1029/2002GC000454.Thornalley, D.J.R., McCave, I.N., Elderfield, H., 2010. Tephra in deglacial ocean

sediments south of Iceland: stratigraphy, geochemistry and oceanic reservoirages. J. Quat. Sci..

Thouveny, N., Bourles, D.L., Saracco, G., Carcaillet, J.T., Bassinot, F., 2008. Paleo-climatic context of geomagnetic dipole lows and excursions in the Brunhes, cluefor an orbital influence on the geodynamo? Earth Planet. Sci. Lett. 275,269e284.

Van Kreveld, S., Sarnthein, M., Erklenkeuser, H., Grootes, P., Jung, S., Nadeau, M.J.,Pflaumann, U., Voelker, A., 2000. Potential links between surging ice sheets,circulation changes and the DansgaardeOeschger cycles in the Irminger Sea,60e18 kyr. Paleoceanography 15 (4), 425e442.

Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K.,Balbon, E., Labracherie, M., 2002. Sea-level and deep water temperaturechanges derived from benthic foraminifera isotopic records. Quaternary Sci.Rev. 21, 295e305.

Weltje, G.J., Tjallingii, R., 2008. Calibration of XRF core scanners for quantitativegeochemical logging of sediment cores: theory and application. Earth Planet.Sci. Lett. 274, 423e438.

Wunsch, C., 2006. Abrupt climate change: an alternative view. Quaternary Res. 65,191e203.

Zielinski, G.A., Mayewski, P.A., Meeker, L.D., Gronvold, K., Germani, M.S.,Whitlow, S., Twickler, M.S., Taylor, K., 1997. Volcanic aerosol records andtephrochronology of the Summit, Greenland, ice cores. J. Geophys. Res. 102,26625e26640.